The objective of this article is to present climate projections for the 2020–2050 period for several climate variables for the State of São Paulo, considering two scenarios RCP 4.5 and RCP 8.5. These projections were based on four climate models (HadGEM2-ES, MIROC5, CanESM2, and BESM) with downscaling of 20 km improved by the CPTEC/INPE Eta Regional Climate Model. Nine variables related to air temperature and rainfall were selected, and the deviation between the observed period (1961–1990) and the projected period (2020–2050) was calculated for the two RCP scenarios for each model and selected variable. Then, the highest and lowest values of the deviations were selected, which represent the maximum and minimum deviations projected for each cell. With this, we obtained the widest range between the possible trends of the four models compared for each variable analyzed in the state of São Paulo. The models used for the two RCP scenarios show the warming of the atmosphere, which tends to be less intense on the shoreline due to the control exerted by the ocean, and greater in the northwest region of the state, furthest from the Atlantic. A reduction in the Cold Spell Duration Index (CSDI) and an increase in the Warm Spell Duration Index (WSDI) is projected, with these trends increasing from south to north of the state. Considering the precipitation, no consensus between the trends of the models used in this study was found, with projections of both increase and decrease in annual total variables and in related extreme indices. Among the different projected trends, the reduction in precipitation is the worst scenario to be faced, given the essential aspect of water. The indicators related to intense precipitation (R95p and RX5day) showed a wide range between the maximum and minimum values projected in the analyzed scenarios, and the increase of these indicators deserves particular attention, since the increase in intense precipitation is a potential trigger of disasters, such as landslides, floods, and accelerated erosion. The results, given their relevance, were applied to the Ecological-Economic Zoning (ZEE) of the State of São Paulo, a technical and political planning instrument that establishes guidelines for land use and management, considering the environmental characteristics and socioeconomic dynamics of different regions of the state.